Controllability

The problem of controllability is becoming more pressing in the wake of the rapid development of increasingly complex assistance systems. The current situation is characterized by increasing severity of controllability issues with rising “efficiency” and “intelligence” of novel assistance systems. In general, the more intensive the support by a system, the harder it is to control in case of error. Of course, one should keep in mind that in many cases, the driver would not be able to avoid a potential accident at all without the system. Thus in assessing driver assistance systems, controllability is strongly coupled to effectiveness; furthermore, effectiveness depends strongly on user acceptance, which ultimately determines take rates and regular use of the system.

Driving and traffic situations requiring parallel processing of multiple information sources within a short time interval pose especially strong challenges for controllability due to the high burden on the driver. Such situations are characteristic of urban areas – as is particularly evident when one considers the systems being developed in the “Cognitive Assistance” project. The aim of the sub-project is thus to develop unified and consistent methodological principles to achieve efficient and valid controllability safeguards, especially in time-critical situations.

These methods and criteria for ensuring controllability will serve to safeguard HMI concepts and functions defined in other UR:BAN sub-projects. UR:BAN functions are being designed to support effective, responsive, accident avoidance. The sub-project “Controllability” will provide established and (where needed) novel methods to allow development and assessment of optimal HMI concepts optimized for sustainable improvement of road safety. The process will take into account limits and potential errors of the systems, possible misuse, as well as override mechanisms.

The sub-project will elaborate procedures for testing and refining system behavior in borderline situations or in case of errors. It will also verify the ability to override the system in normal use. The “Cognitive Assistance” project will provide important applications, but the procedures will not be limited to these applications. Available and novel findings will be used to establish a systematic driver performance knowledge base to support controllability evaluation in various driving and traffic situations. Using this knowledge base, controllability will be assessed in a critical error sequence.